On Tuesday, August 12, 2003, at 01:47 PM, Pete Johnston wrote:
> [Apologies for snipping, but I think these two examples illustrate the
> core of the argument quite nicely]
>
> Eric said:
>
>> Take the example from http://web.resource.org/rss/1.0/modules/dc/
>>
>> <dc:rights>Copyright C 2000 O'Reilly & Associates,
>> Inc.</dc:rights>
>>
>> one may argue this is a "correct" or "incorrect" usage but
>> this example
>> reflects the kinds of information thats associated with the
>> dc:rights
>> element in many examples I see on daily basis.
>>
>> To me (at least) there is not too far of a stretch to represent the
>> unstructured data in the above example in a more explicitly
>> structured
>> way.
>>
>> <dc:rights>
>> <Organization>
>> <name> O'Reilly & Associates, Inc.</name>
>> </Organization>
>> </dc:rights>
>
> I think the distinction between these two examples is that
>
> - in the former the value of the dc:rights property is a statement
> (yes,
> only an unstructured, human-readable statement but a statement all the
> same) that "the copyright in/over the current resource is held by
> O'Reilly Associates". The name of an organisation is part of that
> human-readable statement but it's not the only part.
>
> - in the latter the value of the dc:rights property is a resource of
> type "....#Organization". That's not the same as the first example -
> here, the value is the Organization.
>
> I tend to agree with Andy (in fact I came across the CC usage for the
> first time yesterday and I asked Andy about it, only to find that he
> had
> already raised exactly the same question a few months back): the latter
> seems intuitively, well, a bit odd to me given the current definition
> of
> dc:rights. I agree that "Information about rights held in and over the
> resource" is pretty vague, but it does seem a stretch to say that an
> Organization _is_ "Information about rights held in and over the
> resource".
Hmm... now I'm a bit confused with which problem we're discussing. :)
I thought the argument was with the interpretation / use of dc:rights.
If so, I agree and a better solution (which I alluded to in previous
message) might be to introduce and use a new term 'rightsHolder' whose
value is a resource (e.g. agent). Further, defining this as a
refinement of dc:rights makes sense to me.
Even better would be to leverage DCMI's agent / actors work as the
value for this property but this part of the discussion is perhaps
better left to the agents list.
The other issue you raise (which also a serious one) is with respect to
the value space of the dc:rights property. More specifically weather
the value is a literal or a resource. This problem certainly isn't
isolated to dc:rights and is an issue for all DC elements.
> Not sure this is relevant but FWIW I had a quick look at PRISM earlier
> on today
>
> http://www.prismstandard.org/spec1.2g.pdf
>
> and it seems to me their use of dc:rights is rather closer to what I
> might have expected. They do include a prism:rightsAgent property but
> that is used
>
> - either as a property of the primary resource (i.e. the same resource
> as has the dc:rights property) (see their examples 11 and 13)
>
> - or as a property of a resource that is itself the value of dc:rights
> -
> I think! - this isn't clear as their spec describes things very much
> from the perspective of the XML tree rather than the graph, and there
> isn't an explicit example of this, but I think the last line of the
> "Occurs In" paragraph in 5.3.38 means a use something like:
>
> <rdf:Description>
>
> <dc:title>My resource</dc:title>
> <dc:rights rdf:parseType="Resource">
> <prism:rightsAgent>Someone</prism:rightsAgent>
> [...other properties of the "rights statement"...]
> </dc:rights>
>
> </rdf:Description>
>
> Anyway, in either case PRISM certainly seems to make an effort to avoid
> saying that the value of dc:rights is an Organisation, and that seems
> to
> me more in keeping with the deployments I've seen of dc:rights with an
> unstructured literal value - and rather at odds with the CC usage?
You realize that PRISM description (viewed in RDF terms) is almost
equivalent to the above example? :)
<dc:rights rdf:parseType="Resource">
<prism:rightsAgent>Someone</prism:rightsAgent>
</dc:rights>
is equivalent to
<dc:rights>
<rdf:Description>
<prism:rightsAgent>Someone</prism:rightsAgent>
</rdf:Description>
</dc:rights>
in CC we simply chose to type the resource (Agent rather than Generic)
and used dc:title (which is a mistake!) rather than rightsAgent.
I'm overdue for getting back with Stu to in working on a suggested
solution for this problem... I'm going to refocus some time on this so
apologies in advance if I don't respond in a timely manner to follow-up
messages.
--
eric miller http://www.w3.org/people/em/
semantic web activity lead http://www.w3.org/2001/sw/
w3c world wide web consortium http://www.w3.org/
|